Atos Project
Performance and Feature Importance Report
An overview of this report
This report presents the performance and feature importance of prediction models of four outcomes:
Remission
Past month heroin use
OD
Mortality
Also to address two questions:
- Do you agree with these predictors or should we add more?
## [1] "dg0102" "dg0103" "aust_born"
## [4] "dg0106" "dg0112a" "bpd"
## [7] "trauma" "h0101a" "h0104"
## [10] "dg0117a" "CRIME_1" "dp401"
## [13] "ptsddx" "as01" "sev_dis_pcs01"
## [16] "sev_dis_mcs01" "dailyh_1" "irhprobs1redo"
## [19] "othop_1" "alcohol_1" "cannabis_1"
## [22] "amphet_1" "coke_1" "antidep_1"
## [25] "hall_1" "tobacco_1" "shared_1"
## [28] "unstablehousingBL" "first_high_cat" "first_inj_cat"
## [31] "first_heroin3" "benzos_1" "inhal_1"
## [34] "T1" "T2" "T3"
## [37] "T4"
- This is the list of the columns of HU OD 1 5 10y model.sav
## [1] "id0101" "id0102" "USEDH1_7" "id0701"
## [5] "BL_1YR" "HU_1YR" "HU_5YR" "BL_10YR"
## [9] "HU_10YR" "abs07ons1" "abs07end1" "abs07ons2"
## [13] "abs07end2" "abs07ons3" "abs07end3" "abs07ons4"
## [17] "abs07end4" "abs07ons5" "abs07end5" "abs07ons6"
## [21] "abs07end6" "abs07ons7" "abs07end7" "abs07ons8"
## [25] "abs07end8" "abs07ons9" "abs07end9" "abs07ons10"
## [29] "abs07end10" "abs07ons11" "abs07end11" "abs07ons12"
## [33] "abs07end12" "abs07ons13" "abs07end13" "abs07ons14"
## [37] "abs07end14" "abs07ons15" "abs07end15" "abs07ons16"
## [41] "abs07end16" "abs07ons17" "abs07end17" "abs07ons18"
## [45] "abs07end18" "abs07ons19" "abs07end19" "abs07ons20"
## [49] "abs07end20" "abs07ons21" "abs07end21" "hod0700"
## [53] "hod07epi" "hod07epi10" "hod0701" "hod0702"
## [57] "hod0703" "hod0704" "hod0705" "hod0706"
## [61] "hod0707" "hod0708" "hod0709" "hod0710"
## [65] "hod0711" "hod0712" "hod0713" "BL_YR"
## [69] "BL_MTH" "MonthYear" "INTERVIEWED11YR" "BL_5YR"
## [73] "NDI_2019" "age_death" "COD" "deathtdate"
## [77] "DEATH_YR" "DEATH_MTH" "DEATHDATE_MY" "BL_DMY_1YR"
## [81] "BL_DMY_5YR" "BL_DMY_10YR" "DEATH_1YR" "DEATH_5YR"
## [85] "DEATH_10YR" "BL_DMY_15YR" "DEATH_15YR" "hod_2001"
## [89] "hod_2002" "hod_2003" "hod_2004" "hod_2005"
## [93] "hod_2006" "hod_2007" "hod_2008" "hod_2009"
## [97] "hod_2010" "hod_2011" "hod_2012" "hod_2013"
## [101] "OD_upto1YR" "OD_upto5YR" "OD_upto10YR"
And this is the list of the outcomes
## [1] "STA_1YR" "STA_5YR" "MTA_5YR" "STA_10YR" "MTA_10YR"
## [6] "LTA_10yr" "HU_1YR" "HU_5YR" "HU_10YR" "OD_upto1YR"
## [11] "OD_upto5YR" "OD_upto10YR" "DEATH_1YR" "DEATH_5YR" "DEATH_10YR"
## [16] "DEATH_15YR"
Some are available in the file
## [1] "HU_1YR" "HU_5YR" "HU_10YR" "OD_upto1YR" "OD_upto5YR"
## [6] "OD_upto10YR" "DEATH_1YR" "DEATH_5YR" "DEATH_10YR" "DEATH_15YR"
Some are not available in the file
## [1] "STA_1YR" "STA_5YR" "MTA_5YR" "STA_10YR" "MTA_10YR" "LTA_10yr"
Past month heroin use
HU_1YR
Performance
| .metric | .estimator | .estimate |
|---|---|---|
| accuracy | binary | 0.5900000 |
| f_meas | binary | 0.3692308 |
Feature Importance
HU_5YR
Performance
| .metric | .estimator | .estimate |
|---|---|---|
| accuracy | binary | 0.6300000 |
| f_meas | binary | 0.7730061 |
Feature Importance
HU_10YR
Performance
| .metric | .estimator | .estimate |
|---|---|---|
| accuracy | binary | 0.7300000 |
| f_meas | binary | 0.8439306 |